IFM's K2 Horizon: Fully Open AI Models, from 0.9B to 375B
K2 Horizon: Frontier Performance, Radically Open
IFM releases K2 Horizon, a family of six open models spanning 0.9B to 375B parameters, claiming state-of-the-art performance in the smaller sizes. The release is unprecedented in openness: it includes intermediate checkpoints, training data, code, and logs across the entire training lifecycle, from pretraining through agentic post-training. A key innovation is MoVA (Mixture-of-Value Attention), which applies sparsity to attention, enabling the 36B-A4B model to rival the dense 32B while using far fewer active parameters. All models are Apache 2.0 licensed.
K2 Horizon is the first open model family to expose the complete development process through agentic post-training.
- jjordan
Fully open models really need to be a big part of the AI future. That includes all source code, open training data, how it's organized, fed to the model, processed, etc. Until that becomes a thing you're always going to be left wondering what exactly lies underneath the closed model you are using, leaving open the possibility for societal manipulation.
- a11r
It is great to see another player introduce a fully open stack. Nvidia's Nemotron is the only other prominent one I know of.
All that said, the headline claims do not match the self-reported performance. For example, the dense 32B model is significantly behind Qwen3.8 27B (chart towards the bottom of https://ifm.ai/blog/k2). Gemma4 31B is not in the comparison set. This is the most important sweet spot for self hosted open-weight models today and real competition here will be very welcome.
- kzrdude
The Uno "diffusion adaptor" will take a while for me to understand, but sounds very interesting.
- cogman10
My quick review of the 3.7B model (because I was interested) is that it's not to be trusted for coding.
It failed my basic test I like to ask models and generated incorrect code. When prompted about the bug, it preceded to start hallucinating non-existent APIs. After doing that it got caught in a loop trying to desk check the solution that didn't work.
- cesarvarela
I find it funny that while these releases are a technological miracle, the charts in the doc use tiny fonts and are hard to read. Goes with the idea that coding might be solved, but taste isn't.